3.5. DATA COLLECTION
- In-depth interviews with 17 key informants and 8 Focus group
discussions of 6 members each (2 FGD for unmarried young adults and 6 for
parents) were conducted at appointed and agreed times between 13th
June and 15th July 2005 in Kintampo Town and Ampoma village.
- To minimize information bias, 20 research assistants who
speak both English and Twi were selected and trained on 14th June
(for the 3 assistants involved in qualitative study) and on 25th
June 2005 (for the 17 field workers involved in the survey) in order to enable
them do their work appropriately. These were made of 1 field supervisor, 16
interviewers to administer the structured questionnaire, 1 language translator
for translating the questionnaire, the IDI and FGD-guides and qualitative data
from English into «Twi» (a local dialect), 1 moderator for moderation
of FGDs, and 1 note-taker for note-taking and reporting. All the research
assistants were from KHRC staffs who were assigned to help us in part-term
while still doing their KHRC routine work.
- The pre final questionnaire was pre-tested on
25th June 2005 in Kintampo Sub-District among 20 respondents from
compounds not selected in the study. Six questions were reframed and adjusted
based on the results from feed-back of the pretesting session.
- The survey involved 16 experienced and well trained KHRC
field research assistants. It included at large 170 respondents (150 for study
purpose and 20 for eventual back-up in case of questionnaire disqualification
for incompleteness or inconsistency) and was carried out during 7-10 days from
28th June to 7th July 2005, with average rates of 17-24
questionnaires per day (that is 1-2 questionnaires per day per interviewer) and
30-40 minutes per questionnaire.
- Illustrative photographs on the course of the study were
taken among participants.
3.6. DATA PROCESSING AND ANALYSIS
4.1.11. 3.6.1.
QUALITATIVE DATA
Qualitative data from IDI and FGD were recorded and translated
into English and summarized in Matrix by the researcher. The transcribed
information was reviewed and the main issues summarized.
4.1.12. 3.6.2.
QUANTITATIVE DATA
3.6.2.1. Data quality control
For better quality of data we carefully checked the
completeness and the internal consistency of each questionnaire. Thus out of
the 170 administered questionnaires, twenty found incomplete and with
inconsistent data were simply canceled and replaced by complete ones.
3.6.2.2. Data presentation and statistical
analysis
For easier analysis, pre-coded data from survey questionnaires
were entered into FOXPRO view and then converted into EPIINFO 3.3, STATA and
EXCEL formats that we used in analysis.
Analyzed data was then presented as summarized results in
tables and graphs.
For analytical interpretation, we carried out calculation of
frequencies and relative frequencies. Statistical tests included X2
(chi square), P-values and Odd ratio with 95% confidence intervals. The
level of statistical significance was set at p<0.05.
Two Logit models were used to determine the factors that
affect the probability for a respondent to perceive the need of HIV PCT service
and the probability of willingness to perform premarital HIV counseling and
testing (from Diagram 2). As the data were mostly categorical and the response
binary (Yes/No answers), Logit models were found appropriate to be
used.58
3.6.2.3. Score allocation for level of knowledge and
perception towards HIV PCT
In order to evaluate level of knowledge and level of
perception towards HIV PCT with a more measurable scale, we assigned scores to
answers to specific defined questions. Thus all respondents were given each a
total score of level of knowledge and of level of perception towards HIV PCT
(see ANNEX 4 for details on score attribution per specified question). Then the
following tri-polar marking scale was drawn in order to assess the level of
knowledge and of perception of respondents towards HIV PCT:
|
Level of Knowledge on HIV PCT
|
Level of perception towards HIV PCT
|
|
Range score: 0-22 marks
|
Range score: 0-23 marks
|
|
Acceptable minimum/mean score: 11
|
Acceptable minimum/mean score: 11,5
|
|
Marking scale:
1.Poor knowledge: score < 11
2. Good knowledge: score = 11
a. Average good knowledge: score 11-16,5
b. Adequate good knowledge: score > 16,5
|
Marking scale:
1.Negative/lower/bad perception: score < 11,5
2. Positive/higher perception: score = 11,5
a. Average positive perception: score 11,5-17,5
b. Adequate positive perception: score > 17,5
|
| 
